I remember a different apocrypha for why they skipped from 8 to 10. They wanted avoid OS specific code that conditionally activated from the substring "windows 9" but meant for windows 95 and 98. One would imagine any code like that not being quite as helpful a few decades later.
You misread the GP. The versioning skipping from 8 to 9 was because of bad detection code for windows 95/98. The GP is talking about people staying on Windows 7 until Windows 10 came out, skipping Windows 8.
I don't know the details of that. But even if that's the correct way to determine versions, I think there might be some fraction of software that does it the less correct, more obvious way.
I thought it might be to bring Windows in line with Mac OS 10. Seems petty, but I could see a billion dollar company not liking their flagship is on version 8-9 while the competition is on 10.
Not entirely apocrypha. Among the ones we can most easily name and shame from available source files there were early versions of the Java JDK known to have tests exactly like that in low level library code. Presumably Microsoft's famous app compatibility lab found many more that were closed source that they were not allowed to name and shame.
There's also different apocrypha about the numerology aspect that 9 is a very unlucky number in some cultures and commonly skipped in version numbers (similar to but more so than 13 in the US being skipped on many elevators). (Also why it is said other companies like Apple often skip 9 to make it easier to use the same version cross-culturally without cultural taboo mistakes.)
